[PhilipRoman]

For the CRUD app writers - yeah this is bad news. But it was always clear that it was a bubble. If you're actually solving new problems, it works more like code completion.

Unix was written in assembly with the Ed editor. Even Notepad and any modern language probably represents a 100x increase in productivity compared to that - much more than ChatGPT can do. The field of software engineering will only grow from this.

From what I've seen - most places have way more work than available programmers. Jira tickets stay unsolved for years. Maybe with the increased productivity we can actually clear our backlogs one day.

[Szpadel]

Not sure about you, but I also have a huge backlog of projects that don't exist yet and I would like to create them. but usually they're too complex to do over the weekend and I don't have that much time to spare.

I already tried to approach one of those projects but I quickly failed as gpt knowledge was too outdated for the library that I needed to use and I didn't figure out how to "patch" gpt knowledge.

in this case I'm not very familiar with the stack I needed for that project ( and I think future versions of LLM could get better at bridging that gap) but for tasks that I'm more familiar with I noticed a significant increase in productivity